| Backgroud and purpose:Gastric cancer(GC)is a malignant tumor with a five-year survival rate of only 20-35% for patients treated with surgery alone.Currently,5-fluorouracil(5-FU)based adjuvant chemoradiotherapy is a preferred regimen for postsurgery GC.However,due to the heterogeneity of tumor,the survival outcome of 5-FU based adjuvant chemoradiotherapy varies greatly among different GC patients.Thus,it is necessary to develop a prediction signature to evaluate the survival of GC patients after5-FU based adjuvant chemoradiotherapy,so as to provide an important reference for the clinical treatment of gastric cancer patients.Results: In this study,a total of 1,252 gastric cancer samples and 84 paracancerous samples were collected from Chip and RNA-seq platforms.Among them,577 gastric cancer samples and 84 adjacent normal samples were used for training,and 675 gastric cancer samples were used for validation.First,based on the within-sample relative expression orderings(REOs)of gene expression levels,144,170 reversal gene pairs were identified in 361 cancer samples and 84 paracancerous samples.Among these reversal gene pairs,613 gene pairs that were significant related to the overall survival(OS)of 216 GC patients receiving 5-FU based adjuvant chemoradiotherapy were identified.Then,based on greedy algorithm,34 gene pairs were selected from 631 gene pairs as the optimal prediction signature for 5-FU based adjuvant chemoradiotherapy(C-index = 0.754).Finally,the prediction signature was validated in independent datasets,differentially expressed genes were identified in different prognosis groups and functional enrichment analysis was performed.Result: A signature consisting of 34 gene pairs was identified in training data and validated in three independent datasets.The results showed that in GSE26253＿ACRT,the OS of the high and low risk groups was significantly different(log-rank P=0.033,HR=1.8223,95%CI,1.100-2.994),and the survival time of the high risk group was significantly shorter than that of the low risk group.The similar result were shown in GSE62254＿ACRT(log-rank P=0.028,HR=7.430,95%CI,0.910-60.660)and TCGA＿ACRT(log-rank P=0.042,HR=4.212,95%CI,0.978-18.130).At the same time,disease free survival / relapse-free survival(RFS/DFS)of patients were also analyzed in this study,and the results were similar to OS.A multivariate Cox proportional risk regression model was used to correct for clinical information such as sex,age,and cancer stage.The results showed that the predictive efficacy of 34-GPS was independent of other clinical factors.In addition,this study also verified whether the 34-GPS could predict survival of patients receiving other treatment regiments(surgery alone and chemotherapy alone),the result shown that 34-GPS can not be used as a prognostic signature for GC patients receiving surgery,but it has a certain predictive potential in predicting the survival of GC patients treated with 5-FU based chemotherapy.Next,21 common differentially expressed genes were identified between the different risk groups of the three datasets,and the 21 differentially expressed genes were all highly expressed in the high-risk group samples,and functional enrichment analysis showed that these genes were enriched in multiple tumor progression,gastric cancer invasion and metastasis related pathways.21 differential genes were projected into the protein interaction network,and the results showed that these differential genes were associated with 5-FU-based adjuvant chemoradiotherapy resistance.Conclusion: The signature developed in this study can accurately classify GC patients who may benefit from 5-FU based adjuvant chemoradiotherapy,which could aid clinicians in tailoring more effective GC treatments. |
PDF Full Text Request